Taş, Mehmet Bilge Han; Öztepe, Muhammet Fırat

Objective: The performance of models was evaluated using an ablation-based approach employing attention mechanisms and transfer learning for tumor classification in brain magnetic resonance imaging images. Methods: The glioma, meningioma, no tumor, and pituitary classes from the BRISC2025 public dataset were used. ResNet50, DenseNet121, EfficientNetB0, Xception, and InceptionResNetV2 models were tested using a baseline and the squeeze-and-excitation (SE) attention, spatial attention, and convolutional block attention module attention constructs. The models were trained in two phases: frozen backbone training and fine-tuning, with each experiment repeated three times. Performance was evaluated using accuracy, precision, recall, F1-score, area under the curve (AUC), and loss values. Results: The most successful model was EfficientNetB0 + SE Attention. This model achieved an accuracy of 0.9580 ± 0.0062, an F1-score of 0.9580 ± 0.0057, and an AUC of 0.9949 ± 0.0012 on the test set. Class-based results showed particularly high success rates for the no-tumor and pituitary classes. The most challenging class was meningioma. Conclusion: Transfer learning models supported by attention mechanisms have yielded successful results in brain tumor magnetic resonance classification. The EfficientNetB0 + SE Attention model stood out as the most balanced and successful architecture among the evaluated methods.

Bingöl, Koray; Acar, Tunç Batuhan

Objective: To determine whether anatomical nodal distribution patterns of axillary lymph nodes influence the accuracy of breast magnetic resonance imaging (MRI) nodal staging in patients with breast cancer. Methods: This retrospective study included 156 consecutive patients with newly diagnosed breast cancer who underwent contrast-enhanced breast MRI followed by histopathological nodal evaluation. Two experienced breast radiologists, blinded to pathology results, independently reviewed all MRI examinations. The axillary nodal distribution pattern, the presence of interpectoral (Rotter) lymph nodes, the presence of internal mammary lymph nodes, and the clustered nodal pattern were recorded. MRI-based clinical nodal staging was compared with histopathological nodal status. Multivariable logistic regression identified independent predictors of false-negative MRI nodal staging. Results: Histopathological nodal metastases were identified in 58 patients (37.2%). Breast MRI demonstrated a sensitivity of 70.7% [95% confidence interval (CI), 57.3-81.9%] and a specificity of 84.7% (95% CI, 76.0-91.2%) for detecting nodal metastases. False-negative MRI staging was significantly more frequent among patients with a level II/III-dominant nodal distribution than among those with a level I-dominant pattern (33.3% vs. 25%, P=0.02). Patients with interpectoral (Rotter) lymph nodes had a higher false-negative rate than patients without Rotter lymph nodes (50% vs. 20%, P=0.01). Both a level II/III-dominant nodal distribution [adjusted odds ratio (OR), 2.6; 95% CI, 1.1-6.4] and the presence of interpectoral (Rotter) lymph nodes (adjusted OR, 2.9; 95% CI, 1.2-7.1) independently predicted false-negative MRI nodal staging. Conclusion: Anatomical distribution patterns of axillary lymph nodes influence the accuracy of MRI nodal staging. Assessing nodal distribution, particularly a level II/III-dominant pattern and interpectoral (Rotter) lymph nodes, may improve nodal evaluation and reduce false-negative rates in MRI staging.

Kızılgöz, Volkan; Memis, Kemal Bugra

Objective: Chondromalacia patellae (CP) is a common cause of anterior knee pain and is considered a multifactorial disorder involving mechanical, anatomical, and biomechanical factors. This study aimed to evaluate magnetic resonance imaging (MRI)-derived patellar morphological parameters in patients with and without CP. Methods: This retrospective single-center study included 889 patients who underwent knee MRI from January 2024 to May 2026. Patients were divided into two groups according to the presence or absence of MRI findings compatible with CP. The Wiberg classification, Wiberg index, Wiberg angle, and patellar linear dimensions were evaluated. Results: CP was detected in 278 patients (31.3%), while 611 patients (68.7%) constituted the non-CP group. No significant difference was observed in gender distribution between groups (P =0.151). The CP group was significantly younger than the non-CP group (40.15 ± 10.8 vs. 45.8 ± 12.4 years, P <0.001). The distribution of Wiberg patellar types differed significantly between groups (P=0.021), and type III morphology was more frequent in patients with CP. The Wiberg index was significantly lower in the CP group compared with the non-CP group (0.61 ± 0.08 vs. 0.73 ± 0.07, P<0.001). The anteroposterior diameter was significantly larger (21.18 ± 2.3 mm vs. 19.74 ± 2.1 mm, P <0.001), whereas the transverse diameter was significantly smaller in the CP group (40.33 ± 3.4 mm vs. 41.68 ± 2.9 mm, P <0.001). No significant difference in craniocaudal diameter was found between groups (P=0.137). Conclusion: Several MRI-derived patellar morphological parameters were independently associated with CP after adjustment for age and sex. These findings provide additional information regarding patellar morphology in patients with CP; however, further prospective studies incorporating clinical symptoms and outcomes are required to determine their value in routine MRI interpretation

Oztepe, Muhammet Fırat; Kızılgöz, Berihat

Objective: The purpose of this research was verifying the mean values and ranges of powers ratio (PR), atlanto-dental interval (ADI), basion-dens interval (BDI), and basion-axial interval (BAI) which are critical measurements to evaluate atlantooccipital and atlantoaxial dissociations which indicate instability of the injuries in these critical regions. Methods: After exclusions, 184 female, 87 male, a total of 271 patients’ (the mean age was 48.44 ± 14.99) cervical magnetic resonance images were re-assessed regarding the craniocervical measurements. The mean values and SDs of all parameters were presented. All data distribution results of each ratio and interval measurements were presented with boxplot graphics. Each parameter was analyzed regarding gender differences and age. Results: The mean values of PR, ADI, BDI, and BAI were 0.76 ± 0.07, 1.14 ± 0.61, 5.12 ± 1.74, and 7.41 ± 2.20 respectively. Female and male groups indicated a significant difference regarding ADI and BDI measurements. A significant and negative correlation with ADI and BDI was found regarding age. Conclusion: Besides the mean values and ranges of the 4 critically important parameters, males had higher ADI and BDI values than females, and age showed a negative significant correlation regarding ADI and BDI measurements according to the results of this investigation. More studies and data from larger series need to be added to the literature to create accurate and valuable meta-analyses that will reveal the normal ranges of these critical values and the effect of age and gender on these parameters.

Babatürk, Andelib; Emer, M.Özdeş; Bilezikçi, Banu; Seyfettin, Ayça; Eskalen, Zeynep; Büyükceran, Emre Utkan; Memis, Kemal Bugra; Gültekin, Serap

Objective: To perform a direct, paired comparison of breast magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) for preoperative axillary lymph node staging in patients with breast cancer. Methods: In this retrospective, single-center study, female patients with histologically confirmed breast cancer who underwent both breast MRI and PET/CT prior to surgery were included. Histopathological findings obtained before neoadjuvant therapy served as the reference standard. MRI assessments were based on predefined morphological and functional criteria, whereas PET/CT interpretation relied on focal 18F-fluorodeoxyglucose uptake. Imaging findings were dichotomized on a per-patient basis. Diagnostic performance metrics were calculated, and paired comparisons were conducted using McNemar’s exact test. Results: Among 104 patients, 58 (55.8%) had histopathologically confirmed axillary metastasis. MRI demonstrated a significantly higher sensitivity than PET/CT (96.6% vs. 68.6%) and superior overall diagnostic accuracy (92.3% vs. 81.7%). In contrast, PET/CT exhibited higher specificity (98.1% vs. 87.0%). Agreement with histopathology was very good for MRI (κ=0.83) and good for PET/CT (κ=0.66). A paired analysis showed that all PET-positive cases were also identified by MRI, whereas MRI detected additional metastases not identified by PET/CT (P < 0.001). Conclusion: Breast MRI provides superior sensitivity and overall diagnostic performance, whereas PET/CT offers excellent specificity in axillary staging. Although neither of the modalities replaces surgical nodal assessment, their distinct diagnostic characteristics may inform individualized preoperative risk stratification and multidisciplinary management strategies.